Rolling Again

[Kintaro]

I had to break out the install CD and start from scratch.

Well, that was fun.  Fortunately, I didn't have any critical data on there at all.

Actually, you didn't need to do this, you never do with Linux. What you needed to do is force all those KDE related packages into reinstalling. yum is just a clone of apt-get, I suggest next time you type 'man yum'

the command in english would look like this...

"Yum, my dear pal, please FORCE the base package of KDE and all its dependencies to reinstall"
"Yes good sir."

from memory rpm -f /path/to/rpms/on/media/*.rpm will force a reinstall of everything if you do it from single user mode (init 1) after it you should be able to 'init 5' and have your shit back together without even rebooting.

You can also force remove packages so it doesn't take out their dependencies. For example, you might decide to build it from source to take part in a beta or whatever, same with Firefail (Firefox), if you want to get in on that community spirit thing.

With reinstalls however, and switching distros, because all the apps are virtually identical you should ALWAYS ALWAYS ALWAYS keep /home separate. That way, you can go "fuck it, I want debian" and install debian, tell its paritioner where home is and mark the "dont fucking format" button. This button being there goes all the way back to the damn early naugties and is always there except in Ubuntu's moron mode. Cheers not using that disaster!

[worker201]

The main reason I didn't do at least some of ^that is I couldn't figure out how to get there.  I couldn't launch a terminal window to execute commands, and I had no idea how to drop the runlevel without a terminal.  I can see now that I could have just logged out and gotten a console login.  Oddly, the easiest thing to do was the one I wasn't familiar with.  Lesson learned.

[worker201]

This SELinux stuff can be kinda fun.  LAMP/LAPP stuff is disabled by default.  To get my development environment set up, I had to run:

Code: [Select]

setsebool -P httpd_can_network_connect_db on
I found out about this because the SELinux adviser pops up to tell you how to fix it.  There's a button on the adviser window to "Update Policy", which I figured caused the proposed change to be propagated.  Actually, it turns out that button causes yum to open and attempt to download the latest SELinux policy.  However, yum tries to do this on port 7822 - which causes another SELinux exception, because 7822 isn't approved for yum usage.  Also, the denial of the port causes python to crash, which in turn causes the SELinux adviser to crash.  Good times.  I was able to approve port 7822 for yum usage with:

Code: [Select]

semanage port -a -t http_port_t -p tcp 7822
Back in days of yore, when I was using FCx (x <= 6), I immediately ripped out all SELinux features, because it made configuration so difficult (and troubleshooting impossible).  But these days, it's important to play with policies and configuration, so I'm working with it instead of around it.  It's not like this machine is doing mission critical work, so I feel comfortable fucking around with things.

[worker201]

[Lead Head]

Looks nifty, hows it handling the Intel graphics? I've heard of some issues with various flavors of *nix and Intel chipsets.

[worker201]

I specifically chose this graphics card (and every other part of this computer) because it was 100% supported by Linux.  At the time that I bought it, it was necessary to download drivers for it, but nowdays, it's supported out of the box.

[worker201]

Oh boy, wireless networking.  I didn't have a wireless network at home, so i never had to worry about configuration.  But now that I'm on the road, I have to figure out how to get this going.  Fedora recognizes my card, which is an Intel Pro 3945ABG.  The package that adds the firmware into the kernel seems to be pre-installed, but there must be some sort of configuration step.  I've tried Googling for this info, but I keep getting results that are relevant to Fedora 7 or 8, from before the driver code was merged into the kernel source.

Here's the output from iwconfig, if it means anything to you.

Code: [Select]

lo no wireless extensions

eth0 no wireless extensions

irda0 no wireless extensions

wmaster0 no wireless extensions

wlan0 IEEE 802.11abg ESSID:" "
Mode: Managed Frequency:2.412 GHz Access Point: Not-Associated
Tx-Power=off
Retry long limit:7 RTS thr:off Fragment thr:off
Power Management:off
Link Quality:0 Signal Level:0 Noise Level:0
Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
Tx excessive retries:0 Invalid misc: 0 Missed beacon:0
Assistance much appreciated.  This isn't a priority issue - it is my second computer, after all.

[Kintaro]

Well, NetworkManager should do the hard work for you but barring that, here is the hard way.

Configure /etc/wpa_supplicant.conf so it looks a bit like this...

Code: [Select]

ctrl_interface=/var/run/wpa_supplicant

network={
ssid="Acropolis"
scan_ssid=1
key_mgmt=WPA-PSK
psk="XXXXXXXX"
}

And run it like this
# wpa_supplicant -B -Dwext -iwlan0 -c/etc/wpa_supplicant.conf
the -B switch runs it as a daemon in the background, to get your IP for the interface....
# dhclient wlan0